Pergamon

0361-9230(95)00051-8

Brain Research Bulletin, Vol. 37, No. 6, pp. 605-610, 1995 Copyright © 1995 Elsevier Science Ltd Printed in the USA. All rights reserved

0361-9230/95 $9.50 + .00

Interleukin-lfl Induced Corticosterone Elevation and Hypothalamic NE Depletion Is Vagally Mediated MONIKA FLESHNER, .1 LISA E. GOEHLER,*] ~ JEREMY HERMANN,* JANE K. RELTON,:~

STEVEN F. MAIER* AND LINDA R. WATKINS*

*Department of Psychology, University of Colorado at Boulder, Campus Box 345, Boulder, CO 80309 ?Health Sciences Center, University of Colorado

atSynergen, Inc., Boulder CO

[Received 14 December 1994; Accepted 8 February 1995]

ABSTRACT: Prooesses occurring within the immune system can alter neural funclMn. Cytokines released by calls of the immune system during illness are key messengers in immune-to-brain communication. Intedeukin.l~ (IL-I~) is particularly important in this regard and is known to stimulate a myriad of illness-re- lated outcomes such as fever, sickness behavior, aphagis, adip- sis, hypothalamic-pituitary-adrenal activaUon, and changes in pain reacUv~. Thus peripherally released IL-1/3 has potent neu- ral effects and is a ~ 1 mediator of the impact of immune processes on brain. There is, however, uncertainty conceming the communication pathways involved. We provide evidence that a primary route of peripheral cytokine signaiUng is through s t i m ~ of peripheral vagal all.rents rather than or in addi- tion to direct cytokine access to brain. Subdiaphragmatic, but not hepatic vagotomy, Mocked rhlL-l~-induced hypothalamic norepinephrine depletion and attenuated rhlL-l~-induced in- c r e a m in serum oorticoaterone. These data suggest that rhlL- 1/3 activates the hypothalamic-pituitary-adrenal axis via sum- ulatlon of p a ~ v ~ a i afforonta and further support the hypothe~s that peripheral Gytokine signalling to the CNS is me- diated primarily by stimuist~on of peripheral afferents.

KEY WORDS: Intedeukin-lp, Interisukin-1 receptor antagonist, Illness, HPA, Corticosterone, Norepinephrine, Hypothalamus, Vagotomy, Rat.

INTRODUCTION

There is a great deal of evidence that processes occurring within the immune system are capable of altering neural function [19,23]. Immune system to central nervous system communica- tion was first suggested by Besedovsky and others who reported the discovery that pituitary-adrenal activation occurs at the peak of immune responding to antigen, and that this activation was at least partially mediated by action at the hypothalamus [for re- views see 3, 5]. Indeed, regionally specific electrical and neuro- chemical changes can be measured during the course of a normal immune response [4,9,27,29].

Cytokines released by immune cells during the course of an immune response are key messengers in immune-to-brain sig- nalling [19]. Interleukin-1 (IL-1) is particularly important in this

1 TO whom requests for reprints should be addressed.

regard. Administration of antibody directed against IL-1 or an antagonist directed against the IL-1 receptor (IL-Ira) block many of the neural effects of immune activation [8,20,23]. Further- more, the peripheral administration of IL-1B produces large changes in neural electrical [28] and chemical activity [13], as well as behavioral and physiological sequelae such as fever, aphagia, adipsia, decreased social interaction, reduced explora- tion, and hyperalgesia [19,20,23]. Finally, these same outcomes can be produced by agents such as lipopolysaccharide (LPS) that activate macrophages and lead them to synthesize and release IL- l, and many of the LPS-induced effects can be reduced or blocked by IL-1 immunoneutralization or receptor antagonists [19,20,23].

Although it is clear that peripherally released IL-1 has potent neural effects and is a critical mediator of the impact of immune processes on brain, there is uncertainty concerning the commu- nication pathway(s) involved. The most obvious possibility is that IL-1 crosses the blood-brain barrier and acts on IL-1 recep- tors present in brain [26]. However, IL-1 is a large lipophobic molecule and is therefore unlikely to readily gain entry to the brain through simple diffusion. A number of mechanisms have been proposed that would allow IL-1 to cross into brain. Active transport [2] and penetration at regions in which the barrier is weak or absent [7] are examples.

However, we [31,32,33] have recently suggested an alterna- tive to direct IL-I entry into brain. Peripherally released neu- roactive substances are often capable of activating peripheral af- ferent nerves, and we reasoned that this might be true for IL-lfl. Receptors for IL-lf l are widely distributed in the periphery. The vagus contains afferent fibers that communicate visceral infor- mation to brain [1,22] and vagal electrical activity has been re- ported to increase following injection of IL-1/~ into the hepatic portal vein [25]. We reasoned, therefore, that peripheral IL-lf l might be able to activate vagal afferents, thereby communicating to brain without actual entry. Indeed, peripheral administration of LPS (which causes IL-lf l to be released) induces intense c- fos activation in the nucleus tractus solitarius [15], the major site of vagal afferent termination [ 1,22].

Our initial series of studies explored the hyperalgesic reaction produced by peripheral injection of either IL-lf l or LPS. Intra-

605

606 FLESHNER ET AL.

peritoneal administration of either IL- l/3 or LPS produces a large and sustained enhancement of pain reactivity on both tail-flick to radiant heat and paw-formalin tests [23,32,33,34]. The hyperal- gesia produced by either agent is completely blocked by periph- eral administration of IL-lra [23,34], supporting the notion that the hyperalgesia is mediated by IL-I receptors in the periphery. It is important to recognize that release of IL- 1/3 by macrophages does not occur until several hours following contact with LPS because macrophages do not express IL-1/3 constitutively and so IL-1/3 must first be synthesized before release [24]. It would therefore appear to be problematic that hyperalgesia occurs within 5-10 min following intraperitoneal (IP) injection of LPS. However, specialized macrophage-like cells in the liver called Kupffer cells are able to secrete IL-1 rapidly [11] and intraperi- toneally administered LPS would make a "first pass" through the liver. In addition, liver sinusoidal endothelia produce and release IL-1 during antigenic challenge such as tumor cell inva- sion [30] and the liver has a relatively high density of IL-1 re- ceptors [17,21], some of which appear to be on hepatic nerves [17]. In any case, the foregoing suggested an examination of the role of hepatic vagal afferents, and selective hepatic vagotomy blocked the hyperalgesia produced by LPS, as did a full subdia- phragmatic vagotomy [32].

As already noted, administration of IL-1/3 and LPS-induced release of IL-1/3 produce a variety of outcomes other than hy- peralgesia and it is important to determine which are potentially mediated by activation of vagal afferents. In an initial extension, we found that the hyperthermia that is produced by IL-1 is sub- stantially reduced by subdiaphragmatic, but not by hepatic va- gotomy [31]. Stimulation of adrenal corticosteroid release [6] and depletion of hypothalamic norepinephrine [NE) [13] are among the best substantiated sequelae of peripheral IL-IB administra- tion. Furthermore, the adrenal response to IL-1/3 is mediated, at least in part, by alterations in hypothalamic outflow to the pitu- itary [6]. The purpose of the present experiments was to deter- mine whether subdiaphragmatic or hepatic vagotomy might block or reduce these effects of IL-I.

METHODS

Subjects

Adult male viral-free Sprague-Dawley rats (Sasco; 300- 350 g) individually housed at 25°C with a 12:12 light:dark cycle (lights on at 0600 h) were used in all experiments. Standard rat chow and water was freely available. There were 4 -5 rats per experimental condition. Exact group numbers for each experi- ment are reported in the figure captions. Care and use of the animals were in accordance with protocols approved by the University of Colorado Institutional Animal Care and Use Committee.

Drugs

Recombinant human IL-lfl (rhlL-1/3; Synergen; structurally and functionally identical to IL-1/3) was diluted in sterile saline immediately prior to IP injection at a dose of either 1 #g/kg or 10 #g/kg. Control animals were injected with an equal volume (2 ml/kg) of saline. In Experiment 2, immediately prior to the IP injection of rhlL-1/3, rats received an injection of either rhlL-1/3 receptor antagonist (rhlL-Ira; Synergen) subcutaneously (SC) at a dose of 100 mg/kg of CSE buffer (10 mM citrate, 140 mM sodium chloride and 0.5 mM EDTA, pH 7) or an equal volume (1 ml/kg) of CSE buffer.

Surgeo'

Rats were anesthestized with sodium pentobarbital (55 mg/kg IP) and received either sham surgery, transection of the hepatic branch of the vagus nerve, or complete subdiaphragmatic vagot- omy (including hepatic branch transection). Testing was delayed until 1-2 wks after surgery. See Watkins et al. [32] for surgical and post operative care details.

Blood Sampling and Corticosterone Assessment

One hour, 3 h, and 5 h after rhlL-IB or saline injection, rats were removed from their cages and blood samples (150 #1) were taken from the tall vein. Sampling was completed within 2 min of touching the cage to avoid elevated basal corticosterone. Blood samples were allowed to clot, serum was removed, and frozen at -20°C until later analysis. Serum levels of total corticosterone were measured using a modification of radioimmunoassay (RIA) procedures described by Keith et al. [18] (intra and inter assay CVs < 10%). Corticosterone antibody was purchased from Sigma Chemical Co. (#C-8784).

Hypothalamic NE Assessment

Following the final blood sampling, rats were sacrificed by decapitation and the hypothalamus was quickly dissected over dry ice, weighed, and frozen at -70°C. Hypothalamic NE levels were measured using high performance liquid chromatography (HPLC) as previously described [12].

Statistical Analysis

Repeated measure and factorial ANOVAs were performed on the NE data and the corticosterone data respectively. Fischer Pro- tected Least Significance Difference (F-PLSD) tests were per- formed post hoc.

RESULTS

We first sought to determine a dose of rhlL-l/3 that would reliably elevate serum corticosterone and deplete hypothalamic NE in sham operated rats. As can be seen in Fig. 1A, IP rhlL-1/3 resulted in a dose dependent depletion of hypothalamic NE 5 h after injection (F(2, 13) = 12.5, p < .001) with a reliable deple- tion occurring in rats injected with 10 #g/kg of rhlL-lfl (10 #g/ kg of rhlL-1/3 vs. saline; F-PLSD = 163.9, p < .05). As evident in Fig. 1B, serum corticosterone levels also showed a dose de- pendent elevation and a dose dependent duration of increase. These observations are supported by a reliable main effect of drug (F(2, 15) = 44.2, p < .0001) and a reliable drug by time of blood sample interaction (F(4, 30) = 9.9, p < .0001). One/zg/ kg of rhlL- 1/3 reliably elevated serum corticosterone levels above basal levels I h (F-PLSD = 10.4, p < .05) and 5 h (F-PLSD = 7.7, p < .05) but not 3 h (F-PLSD = 6.6, p > .05) after injection. In contrast, 10 #g/kg of rhlL-1B resulted in elevated corticoste- rone levels 1 h, 3 h, and 5 h after injection (p < .05).

We next sought to verify the receptor specificity of the effect of rhlL-1/3 on corticosterone release and hypothalamic NE de- pletion. Rats were injected SC with 100 #g/kg rhlL-lra or equi- volume vehicle. Immediately after the first injection, rats re- ceived 10 #g/kg of rhlL-1/3 or equivolume vehicle IP. RhlL-lra blocked the hypothalamic NE depletion produced by 10 #g/kg of rhlL-1/3 (Fig. 2A). This observation is supported by a reliable drug (rhlL-I ra or vehicle) by drug (rhlL-1/3 or saline) interaction (F(I, 15) = 7.1, p < .01). RhlL-lra also blocked the rhlL-l/3- induced increase in serum corticosterone (Fig. 2B). This obser-

VAGOTOMY ATTENUATES IP IL-1/3 ACTIONS 607

AO 1600

lSm

!- 0 1 10

Dose of rhIL-113 (ttg/kg)

levels in subdiaphragmatic rats as compared to saline subdia- phragmatic controls at 1 h (p > .05) and 3 h (/7 > .05) but did reliably increase it 5 h (F-PLSD = 10.1, p < .05) after rhIL-1/3. Subdiaphragmatic vagotomy also resulted in a reliable reduction in rhIL-1/3-induced elevation of serum corticosterone compared to rhIL-1/3-injected sham operated rats at 1 h, 3 h, and 5 h after injection (p < .05 for each comparison).

We then sought to investigate the effect of vagotomy when restricted to the hepatic branch. Selective hepatic vagotomy did not block the rhIL-1/3-induced depletion of hypothalamic NE (Fig. 4A). This is supported by a reliable main effect of drug

40 B.

Timecourse (Hrs)

FIG. 1. RhlL-1/~ produces dose-dependent changes in hypothalamic nor- epinephrine (NE) content and serum corticosterone levels. (A) Hypotha- lamic NE (expressed as ng NE per g of tissue) was reliably decreased by 10, but not 1, #g/kg rhlL-l/~. (B) Serum corticosterone levels (/zg corticosterone per dl serum) were transiently increased by 1 #g/kg rhlL- 1/~ (filled circles) and showed sustained elevation following 10 #g/kg rhIL-l/~ (filled squares), compared to vehicle controls (open triangles).

vation is also supported by a reliable drug (rhlL-lra or vehicle) by drug (rhlL- 1/~ or saline) interaction (F(1, 15) --- 6.6, p < .025).

The effect of sudiaphragmatic vagotomy on the rhlL-1/~-in- duced hypothalarnic NE depletion as shown in Fig. 3A. Section- ing the vagus just below the diaphragm (including the hepatic branch) blocked the hypothalarnic NE depletion induced by pe- ripheral injection of 10 #g/kg of rhIL-1/3. Although there was no reliable overall interaction (F(1, 20) = 1.9, p > .05) due to an effect of subdiaphragnmtic surgery alone on hypothalamic NE (F(1, 20) = 6.1, p < .05), post hoe analysis revealed a reliable effect of rhlL-1]~ in sham operated rats (F-PLSD = 404.4, p < .05) and no reliable effect of rhlL-1/~ in subdiaphragmatic op- erated rats (F-PLSD = 112.3, p > .05). The blockade of adrenal output was less complete. As seen in Fig. 3B, sudiaphragmatic vagotomy attenuated the rhlL-1/3-induced increase in serum cor- ticosterone. This observation is supported by a reliable drug by surgery interaction (F(1, 20) = 5.3, p < .05). Post hoc analyses reveal that rhIL-1/3 did not reliably increase serum corticosterone

Ao 2000-

B rUlL-tl5 (10 ~g~S) Vehide

0 100 Dose of rhIL.lra (mg/kg)

BQ 40 • rhlL-lr~rE~-l~

0 VehbhlL-ll5 • rhlL-lrsNeh [] Ve~V~

.i ,i 0

i i i Timecourse (Hrs)

FIG. 2. RhIL-1/3 induced changes in hypothalamic norepinephrine (NE) content and serum corticosterone levels are blocked by an IL- 1 receptor antagonist. (A) Compared to vehicle injected controls, hypothalamic NE (expressed as ng NE per g of tissue) was reliably decreased by 10 pg/kg rhIL-1/~, when rats received only CSE vehicle (that is, received 0 pg/kg rhIL-lra). This effect on hypothalamic NE content was blocked in rats which received rhIL-lra (100/~g/kg) immediately prior to rhIL-l/L (B) In rats that received no rhlL-lra, serum corticosterone levels (#g corti- costerone per dl serum) were reliably increased by 10 pg/kg rhIL-1/3 (open circles), compared to vehicle controls (open squares). This effect on serum cortieosterone was blocked in rats which received 100 ~g/kg rhIL-lra immediately prior to rhlL-1/~ (filled circles). RhIL-lra had no effect on corticosterone levels, in the absence of rhIL-1/3 (filled squares).

608 FLESHNER ET AL.

2250,

i

AO

I I r " - I rhlL-11~ (10 ;tg/kg)

Vehicle

I

injection of an IL-1 receptor antagonist, supporting the conclu- sion that these effects are due to rhlL- 1/3 acting at IL- 1/3 receptors and not via nonspecific drug action. These findings replicate in the rat those reported by others in the mouse [13].

Subdiaphragmatic vagotomy but not hepatic vagotomy blocked the IL-1/3-induced depletion of hypothalamic NE. This finding suggests that a great deal of the peripheral signal of cy- tokines to the brain is conveyed via the vagus and must include vagal branches other than those innervating the liver. This finding is especially provocative given the relatively large dose of IL-I/3 used. Clearly the larger the dose the greater the opportunity for direct cytokine entry into brain from the general circulation. Nev-

Subdiaph. Sham Vagotomy . ~

B. • rhlL-l~/Subdia. Vagot. O Veh/Subdia. Vagot. • rhl~l. ~/Slmm 1:3 Veh/Sham

i 3 Timecourse (Hrs)

" " 3O

20 e~ O tm ~ 10

o

0

FIG. 3. RhlL-1/3 induced changes in hypothalamic norepinephrine (NE) content and serum corticosterone levels are attenuated by subdiaphrag- matic vagotomy. (A) Compared to vehicle injected controls, hypotha- lamic NE (expressed as ng NE per g of tissue) was reliably decreased by 10 #g/kg rhlL-1/3, in sham operated animals. This effect of rhlL-I/3 on hypothalamic NE content was blocked by subdiaphragmatic vagot- omy. (B) In sham operated controls, serum corticosterone levels (/~g corticosterone per dl serum) were reliably increased by 10 #g/kg rhIL- 1/3 (filled squares), compared to vehicle controls (open squares). This rhlL-1/3 effect on serum corticosterone was attenuated by subdiaphrag- matic vagotomy prior to rhlL-l/3 (filled circles), compared to vehicle injected vagotomy controls (open circles).

(F(1, 16) = 6.6, p < .025) but no reliable drug by surgery inter- action (p = .64). Figure 4B also clearly shows no effect of hepatic vagotomy on the rhlL-1/3-induced increase in serum corticoste- rune. This is supported by a reliable main effect of drug (F(I, 16) = 18.4, p < .001) but no drug by surgery interaction (p > .05).

DISCUSSION

RhlL-1/3 at a dose of 10 /~g/kg resulted in a hypothalamic depletion of NE and elevated serum corticosterone as long as 5 h after a single IP injection. This effect is totally blocked by prior

1200¸

1000,

AO mm run,.ll~ (lo ~g/kg) 8---7 Vehlde

T

Hepatic Vagotomy

Sham

BO • rhIL.l~lepatic Vagot. A Veh/Hepatlc Vagot. • rhlL.l~lmm 1"3 Veh/Sham

50

10

0

Tlmecourse (Hrs)

FIG. 4. RhIL-1/3 induced changes in hypothalamic norepinephrine (NE) content and serum corticosterone levels are not attenuated by hepatic vagotomy. (A) Compared to vehicle injected controls, hypothalamic NE (expressed as ng NE per g of tissue) was decreased by 10 #g/kg rhlL- 1/3, in both sham operated and hepatic vagotomized animals. (B) In sham operated controls, serum corticosterone levels (#g corticosterone per dl serum) were reliably increased by 10 #g/kg rhlL-lfl (filled squares), compared to vehicle controls (open squares). This rhlL-1/3 effect on se- rum corticosterone was not reliably attenuated by hepatic vagotomy prior to rhlL-1/3 (filled triangles), compared to vehicle injected vagotomy con- trols (open triangles).

VAGOTOMY ATTENUATES IP IL- I~ ACTIONS 609

ertheless, subdiaphragmatic transection of the vagus blocked NE depletion. The fact that subdiaphragmatic vagotomy blocked the NE depletion produced by peripheral administration of IL-1/3 suggests that either IL-I/3 does not enter the brain or that hypo- thalamic NE depletion is not mediated by rhlL-1/3 acting in the brain. It would appear that the former of these two possibilities is true and NE depletion in response to peripheral rhlL-1/3 is mediated by peripheral stimulation of vagal input. IL-1/~ may enter the brain from the circulation but this does not seem to cause the depletion of hypothalamic NE.

Subdiaphragmatic vagotomy, but not hepatic vagotomy, at- tenuated but did not completely block the IL-1B-induced increase in serum corticosterone. Thus again, at least part of the signal from 11_,-1/3 to brain is carried by the vagus. There are at least three possible explanations for the corticosteroid response that remained after vagotomy. One is that IL-1/3 entry into brain was sufficient to initiate a corticosterone response, even though it was not sufficient to deplete hypothalamic NE. This possibility is con- sistent with the fact that a dose of 1 #g/kg of rhlL-1B produced an increase in serum corticosterone but did not lead to a depletion of NE. The 10 #g/kg dose was required to deplete hypothalamic NE. Thus a small amount of IL- I~ entry into brain may activate areas in the hypothalamus such as the paraventricular nucleus, but be unable to produce a sustained enough response to lead to depletion. The second possibility is that the peripherally admin- istered IL-1/3 acted directly at IL-1/3 receptors on the pituitary or adrenal to produce part of the corticosteroid response. A variety of evidence suggests that IL- I~ and LPS can indeed exert this sort of direct action at the level of the pituitary and adrenal [for review see 10,14], and so vagotomy may have reduced but not eliminated the corticosteroid increase because it eliminated the central hypothalamic component of the pathway between IL-1B and corticosteroid response but not the peripheral component. The final possibility is that there are other peripheral afferents that can be activated by IL-1/3 or that vagal branches beyond the diaphragm are involved. The fact that IL-1 ra completely blocked the corticosteroid response is consistent with either of these hy- potheses.

We have previously reported that selective section of the hepatic branch of the vagus blocks LPS-induced hyperalgesia [32], but has no effect on IL- 1/3-induced fever [31 ]. This could mean that IP LPS stimulated IL-1/3 (most likely predominantly in the liver) and IP IL-1/3 act at different vagal sites or that hyperalgesia and fever are mediated by different vagal branches. The fact that hepatic vagotomy had no effect in the present experiments is consistent with the first of these pos- sibilities.

The overall pattern that emerges from the present experi- ments and prior work [23,31,32,33,34] suggests that a great deal of IL - I~ signalling to the CNS is accomplished by acti- vation of vagal afferents rather than by direct access to brain. IL-1/3 produced hyperalgesia [33], fever [31], condit ioned taste aversions [16], and now hypothalamic NE depletion and corticosterone increases are all completely or largely elimi- nated by vagotomy. The fact that different afferent branches of the vagus might mediate different aspects of the response to IL - I~ is not unexpected if cytokine release is viewed as a local event as well as a long range hormonal signal. It should be emphasized, however, that there is no incompatibil i ty be- tween different reactions to IL-1/3 being differentially medi- ated by the various signall ing sources. It will be of interest to determine whether other neuroactive cytokines such as TNF and IL-6 that are also released during illness and pathogenic challenge communicate to brain in a similar fashion.

ACKNOWLEDGEMENTS

We would like to thank Lee Silbert and Julie Barter for their excellent technical assistance. This work was supported by Synergen, Inc., NIH Grant MH45045 and NIH Grant NS31569.

REFERENCES

1. Adachi, A. Projection of the hepatic vagal nerve in the medulla ob- longata. J. Auton. Nerv. Syst. 10:287-293; 1984.

2. Banks, W. A.; Kastin, A. J.; Durham, D. A. Bidirectional transport of interleukin-1 alpha across the blood-brain barrier. Pept. Brain Res. Bull. 23:433-437; 1989.

3. Berkenbosch, F.; DeFijk, R.; Schotanus, K.; Wolvers, D.; VanDam, A. M. The immune-hypothalamo-pituitary adrenal axis: Its role in im- munoregulation and tolerance to self antigens. In: Rothwell, N.; Dantzer, R., eds., Interleukin-I in the brain. New York: Oxford; 1992:75-92.

4. Berkenbosch, F.; VanOers, J.; DelRey, A.; Tilders, F.; Besedovsky, H. O. Corticotropin releasing factor producing neurons in the rat are activated by interleukin-l. Science 238:524-526; 1987.

5, Besedovsky, H. O.; del Rey, A. E.; Sorkin, E. What do the immune system and the brain know about each other? Immunology Today 4:342-346; 1983.

6. Besedovsky, H. O.; DelRey, A.; Kinsman, I.; Furukawa, H,; Monge Arditi, G.; Kabiersch, A. Cytokines as modulators of the hypothal- amus-pituitary-adrenal axis. J. Steroid Biochem. Mol. Biol. 40:613- 618; 1991.

7. Blatteis, C. M.; Bealer, S.; Hunter, W. S.; Llanos, Q. J.; Ahokas, R. A.; Mashbum, T. A. J. Suppression of fever after lesions of the anteroventral third venuicle in guinea pigs. Brain Res. Bull. 2:519-526; 1983.

8. Bluthe, R.-M.; Dantzer, R.; Kelley, K. W. Effects of interleukin-1 receptor antagonist on the behavioral effects of lipopolysaccharide in the rat. Brain Res. 573:318-320; 1992.

9. Carlson, S. L.; Felton, D. L.; Livnat, S.; Felten, S. Y. Alterations of monoamines in specific central autonomic nuclei following immu- nization. Brain, Behavior & Immunity 1:52-64; 1987.

10. Cunningham, E. T.; DeSouza, E. B. Interleukin 1 receptors in the brain and endocrine tissues. Immunoi. Today 14:171 - 176; 1993.

11. Decker, K. Biologically active products of stimulated liver macro- phages (Kupffer cells). Eur. J. Biochem. 192:245-261; 1990.

12. Desan, P. H.; Woodmansee, W. W.; Ryan, S. M.; Smock, T. K.; Maier, S. F. Monoamine neurotransmitters and metabolites during the estrous cycle, pregnancy, and postpartum period. Pharmacoi. Biochem. Behav. 30:563-568; 1988.

13. Duun, A. J. Endotoxin-induced activation of cerebral catecholamine and serotonin metabolism:comparison with interleukin-l. J. Phar- macol. Exp. Ther. 261:964-969; 1992.

14. Elenkov, I. J.; Kovacs, K.; Kiss, J.; Bertok, L.; Vizi, E. S. Lipopolysac- charide is able to bypass corticotrophin-relasing factor in affecting plasma ACTH and corticosterone levels: Evidence for rats with lesions of the paravenuicular nucleus. J. Endocrinol. 133:231-6; 1992.

15. Ericsson, A.; Kovacs, K. J.; Sawchenko, P. E. A functional anatom- ical analysis of central pathways subserving the effects of interleu- kin-I on stress-related neuroendocrine neurons. J. Neurosci. 14:897-913; 1994.

16. Goehler, L. E.; Busch, C. R.; Tartaglia, N.; Relton, J.; Sisk, D.; Maier, S. F.; Watkins, L. R. Blockade of cytokine induced condi- tioned taste aversion by subdiaphragmatic vagotomy: Further evi- dence for vagal mediation of immune-brain communication. Neu- rosci. Lett. 185:163-166; 1995.

17. Goehler, L. E.; Relton, J.; Maier, S. F.; Watldns, L. R. Biotinylated interleukin-I receptor antagonist (IL-lra) labels paraganglia in the rat liver hilus and hepatic vagus. Proc. Soc. Neurosci. (in press)

18. Keith, L. D.; Winslow, J. R.; Reynolds, R. W. A general procedure for estimation of corticosteroid response in individual rats. Steroids 31:523-531; 1978.

19. Kent, S.; Bluthe, R.-M.; Kelley, K. W.; Dantzer, R. Sickness behav- ior as a new target for drug development. Trends Pharmacol. Sci. 13:24-28; 1992.

20. Kiuger, M. J. Fever: Role of pyrogens and cryogens. Physiol. Rev. 71:93-127; 1991.

610 FLESHNER ET AL.

21. Kohira, T.; Matsumoto, K.; Ichihara, A.; Nakamura, T. Identification of a biologically functional novel IL- 1 beta-specific receptor on adult rat hepatocytes. J. Biochem (Tokyo) 114:658-62; 1993.

22. Leslie, R. A.; Reynolds, J. M.; Lawes, I. N. C. Central connections of the nuclei of the vagus nerve. In: Ritter, R.; Ritter, S. C.; Barnes, C. D., eds. Neuroanatomy and physiology of abdominal vagal af- ferents. Ann Arbor: CRC Press; 1992:81-98.

23. Maier, S. F.; Wiertelak, E. P.; Martin, D.; Watkins, L. R. Interleukin- 1 mediates the behavioral hyperalgesia produced by lithium chloride and endotoxin. Brain Res. 623:321-324; 1993.

24. Nathan, C. F. Secretory products of macrophages. J. Clin. Invest. 79:319-326; 1987.

25. Niijima, A. The afferent discharges from sensors for interleukin-I- beta in hepatoportal system in the anesthetized rat. J. Physiol. 446:236P; 1992.

26. Rothwell, N.; Dantzer, R. lnterleukin-1 in the brain. New York: Ox- ford; 1992.

27. Saphier, D.; Abramsky, O.; Mor, G.; Ovadia, H. Multiunit electrical activity in conscious rats during an immune response. Brain, Behav- ior and Immunity 1:40-51; 1987.

28. Saphier, D.; Ovadia, H. Selective facilitation of putative corticotro- pin-releasing factor-secreting neurones by interleukin-1. Neurosci. Lett. 114:283-288; 1990.

29. Saphier, D.; Ovadia, H.; Abramsky, O. Neural responses to antigenic challenges and immunomodulatory factors. Yale J. Biol. Med. 63:109-119; 1990.

30. Vidal-Vanaclocha, F.; Amezaga, C.; Asumendi, A.; Kaplanski, G.; Dinarello, C. A. Interleukin-1 receptor antagonist (IL-lra) inhibits hepatic metastasis of intrasplenically injected B16 melanoma cells. Proc. Annu. Meet. Am. Assoc. Cancer Res. 34:A431 ; 1993.

31. Watkins, L. R.; Goehler, L. E.; Relton, J. K.; Tartaglia, N.; Silbert, L.; Martin, D.; Maier, S. F. Blockade of interleukin-I induced fever by subdiaphragmatic vagotomy:evidence for vagal mediation of immune-brain communication. Neurosci. Lett. 183: 27-31; 1995.

32. Watkins, L. R.; Wiertelak, E. P.; Goehler, L.; Mooney-Heiberger, K.; Martinez, J.; Furness, L.; Smith, K. P.; Maier, S. F. Neurocir- cuitry of illness-induced hyperalgesia. Brain Res. 639:283-299; 1994.

33. Watkins, L. R.; Wiertelak, E. P.; Goehler, L. E.; Smith, K. P.; Martin, D.; Maier, S. F. Characterization of cytokine-induced hyperalgesia. Brain Res. 654:15-26; 1994.

34. Wiertelak, E. P.; Smith, K. P.; Fumess, L.; Mooney-Heiberger, K.; Mayr, T.; Maier, S. F.; Watkins, L. R. Acute and conditioned hy- peralgesic responses to illness. Pain 56:227-234; 1993.